Professional Documents
Culture Documents
Psychological Bulletin
2008, Vol. 134, No. 5, 700 741
Macquarie University
Cancer-related fatigue (CRF) is a significant clinical problem for more than 10 million adults diagnosed
with cancer each year worldwide. No gold standard treatment presently exists for CRF. To provide a
guide for future research to improve the treatment of CRF, the authors conducted the most comprehensive
combined systematic and meta-analytic review of the literature to date on non-pharmacological (psychosocial and exercise) interventions to ameliorate CRF and associated symptoms (vigor/vitality) in
adults with cancer, based on 119 randomized controlled trials (RCTs) and non-RCT studies. Metaanalyses conducted on 57 RCTs indicated that exercise and psychological interventions provided
reductions in CRF, with no significant differences between these 2 major types of interventions
considered as a whole. Specifically, multimodal exercise and walking programs, restorative approaches,
supportive expressive, and cognitive behavioral psychosocial interventions show promising potential
for ameliorating CRF. The results also suggest that vigor and vitality are distinct phenomena from CRF
with regard to responsiveness to intervention. With improved methodological approaches, further
research in this area may soon provide clinicians with effective strategies for reducing CRF and
enhancing the lives of millions of cancer patients and survivors.
Keywords: cancer-related fatigue, meta-analysis, systematic review, psychological therapies, exercise
Supplemental materials: http://dx.doi.org/10.1037/a0012825.supp
CANCER-RELATED FATIGUE
701
702
et al., 2005; Stevinson, Lawlor, & Fox, 2004). The Markes et al.
(2006), McNeely et al. (2006), and Schmitz et al. (2005) reviews
also included meta-analyses of this literature. A notable limitation
of all five reviews is that CRF was only a partial aim of these
reviews, and the number of studies that qualified for inclusion in
each of the reviews was quite limited. More specifically, Stevinson
et al. (2004) systematically evaluated 25 controlled studies (including both RCTs and nonrandomized controlled clinical trials
[CCTs]), which investigated the effects of physical exercise interventions in promoting general psychological well-being. Only 12
of these trials assessed fatigue outcomes. Similarly, Knols et al.
(2005) examined the methodological quality and efficacy of physical exercise in 34 controlled trials (both RCTs and CCTs) in
managing physical functioning and psychological well-being, but
only 15 of these trials included fatigue as a primary or secondary
outcome measure.
Schmitz et al. (2005) conducted a combined systematic and
meta-analytic review of 32 controlled physical exercise trials in
improving 25 outcome variables, including fatigue/tiredness and
vigor/vitality outcomes. Whereas 9 of the 32 trials assessed fatigue/tiredness outcomes (either during or following cancer treatment) and were found to have a combined small effect size (.14),
three trials assessed vigor/vitality outcomes and were found to
have a moderate effect when the intervention was administered
during cancer treatment (.43) and a large positive effect when the
intervention was administered following cancer treatment (.83).
McNeely et al. (2006) focused their review on evaluating the
effectiveness of 14 physical exercise RCTs in improving quality of
life and physical functioning in breast cancer patients. Six of these
studies included fatigue as an outcome variable, and meta-analyses
revealed a moderate to large positive effect of the intervention
(.72), although only 2 of these trials were evaluated as having high
internal validity. Similarly, in a recent Cochrane review, Markes et
al. (2006) investigated the methodological quality and efficacy of
aerobic and/or resistance exercise interventions for breast cancer
patients who were concurrently undergoing adjuvant cancer treatment. Five out of the 9 trials included in this review used fatigue
as an outcome measure. Interestingly, these researchers found a
statistically nonsignificant effect on fatigue symptoms for exercise
interventions (standard mean difference .12). This result is
contrary to McNeely et al.s findings. The discrepancy between
these two reviews may be due in part to differences in study
inclusion criteria. Whereas McNeely et al. included interventions
that were administered either during or following adjuvant cancer
treatment, Markes et al.s review was explicitly based on evaluating the efficacy of exercise interventions conducted during adjuvant cancer therapy. Indeed, Markes et al.s finding is in part
comparable to the fatigue results reported by Schmitz et al. However, the fatigue results from Schmitz et al.s and McNeely et al.s
meta-analytic findings are substantially different. Whereas McNeely et al. explicitly noted the studies that were used in the
meta-analysis, as did Markes et al., Schmitz et al. did not specify
which studies were included. Consequently, this limits the comparisons that can be made among these three meta-analytic reviews
for determining the extent to which physical exercise interventions
are beneficial in reducing fatigue symptoms across various cancer
patient populations.
Lawrence et al. (2004) published the first systematic review that
specifically evaluated the effectiveness of psychosocial as well as
The Jacobsen et al. (2007) review was published following the completion of the current systematic and meta-analytic review.
CANCER-RELATED FATIGUE
METHOD
Search Strategy
The following electronic databases were searched from their
respective inceptions through to the end of December 2006:
CANCERLIT, CINAHL, EMBASE, MEDLINE, PubMed, and
PsycINFO. The searches were conducted with the following
subject headings and/or keywords and combinations: (a) cancer
(including tumor/tumour, neoplasm/s); (b) fatigue-related key-
703
words (fatigue, vitality, vigor/vigour), tired/ness, energy, inertia, lethargy); (c) cancer-related symptoms (quality of life
[QOL], physical functioning, psychological functioning, mood,
distress, anxiety, depression, stress); (d) non-pharmacological
interventions (counseling/counselling, psychotherapy, stress
management, cognitive behavioral (behavioural) therapy
[CBT], behaviour/behavior therapy, relaxation training, exercise, massage, education); and (e) common psychosocial quantitative measures, which included fatigue-related subscales
(POMS, Functional Assessment of Cancer Therapy [FACT and
FACIT], the Core scale from the European Organization for
Research and Treatment of Cancer Quality of Life Questionnaire [EORTCQLQ-C30], and the MOS-SF-36). The bibliographies of retrieved articles, narrative reviews, and commentary articles on CRF and psychosocial interventions for cancer
patients were also manually searched for additional references.
The abstracts of all articles identified by electronic and manual
searches (4,040 in total) were carefully screened by Maria
Kangas for consideration of inclusion in this review, and a
random selection of 25% of the abstracts were independently
assessed by Guy H. Montgomery to check for consistency in
selection. All abstracts and/or titles of articles that were considered for inclusion were read independently by these two
authors, applying the selection criteria stipulated below.
704
Meta-Analytic Procedure
Fatigue
Of the 67 RCTs that assessed fatigue or tiredness as an outcome
measure (50 psychosocial and 17 exercise interventions) and met
inclusion criteria, 10 studies (9 psychosocial trials and 1 exercise
trial) were excluded because they did not report sufficient statistical information to compute effect sizes. Hence, in total, 85% (k
57) of RCT interventions that met the initial inclusion criteria for
the meta-analysis were included in the meta-analyses.
Vigor/Vitality
Of the 60 RCTs that assessed vigor/vitality as an outcome
measure (50 psychosocial and 10 exercise interventions) that met
initial inclusion criteria, 17 trials (14 psychosocial and 3 exercise
CANCER-RELATED FATIGUE
trials) were excluded from the meta-analysis, as they did not report
sufficient statistical information to compute effect sizes. Thus, in
total, 72% of RCT studies were included in the meta-analyses.
Effect size (d) was calculated separately for (a) fatigue and
tiredness outcome measures (henceforth referred to as fatigue
outcomes and (b) vigor, vitality, and increased energy outcome
measures (henceforth referred to as vigor/vitality outcomes). For
study trials that reported statistical data on two or more measures
of the same outcome (e.g., fatigue), consistent with the systematic
review analyses, the effect size was calculated with the outcome
measure that yielded more favorable findings. This latter criterion
was only applicable for a couple of interventions. For each individual study trial, effect sizes (ds) for each variable of interest (i.e.,
fatigue and vigor/vitality outcomes) were computed by taking the
difference between the control group mean and the experimental
group mean, then dividing by the standard deviation (J. Cohen,
1988). For studies that did not provide this descriptive statistical
information but did report relevant values from statistical tests
(e.g., t test, F test, p values), effect sizes were estimated on the
basis of the formulas in Smith, Glass, and Milles (1980) text
(Appendix 7). In some instances where no relevant statistical
information was reported (including the specific p value), but the
outcome variable (e.g., fatigue) was reported to be significant, the
effect size was derived by assuming that the p values were equivalent to .05. In all other instances where insufficient statistical
information was reported to estimate the group means and/or
standard deviation or mean standard error term, these study trials
were excluded from the meta-analysis.
The meta-analytic computations were performed with the Stata
(version 9) program. To control for the wide variability in sample
size, all calculations were weighted according to sample size.
Weighted mean effect sizes and the 95% confidence intervals for
the weighted means were calculated with a random effect approach. This approach was used for two reasons: (a) the effects for
all of the outcomes were heterogeneous and (b) the random approach is more appropriate when a relatively large number of
studies are to be analyzed. A notable advantage of this method is
that it has a more powerful scope in generalizing to similar studies
not included in the meta-analysis (cf. Frattaroli, 2007). A p value
of less than .05 was considered to be a significant effect size.
705
RESULTS
The findings from the systematic review are presented first,
followed by the results from the meta-analytic component of this
combined review.
Systematic Review
Search Results
A total of 119 published articles comprising 131 studies (with 8
studies that included separate follow-up publications) met the
inclusion criteria. Three groups of researchers published what
appear to be findings from the same databases in two distinct
publications (Allison, Edgar, et al., 2004; Allison, Nicolau, et al.,
2004; Coleman, Coon, et al., 2003; Coleman, Hall-Barrow, 2003;
A. L. Schwartz, 1999, 2000a); accordingly, these articles were
only counted once in the final tally of included studies. Seventyone studies were RCTs (n 52 psychosocial interventions, including 3 separate published follow-up studies, and n 17 physical exercise interventions); 9 were non-RCTs or CCTs (n 6
psychosocial with one separate follow-up paper; n 2 exercise);
and 36 were single-group designs (n 21 psychosocial, including
4 separate published follow-up studies, and n 11 exercise).
However, we excluded two trials conducted by the same researchers (Cimprich, 1993; Cimprich & Ronis, 2001), which examined the
effects of a restorative psychosocial intervention in improving attentional fatigue in breast cancer patients. The reason for excluding these
trials was that the fatigue outcome was based on participants combined performance on several neurocognitive tests (including Digit
Span Forward and Backward from the Weschler Intelligence Assessment Scale [WAIS)]). Therefore, we deemed that these two studies
substantially deviated from the conventional unidimensional and multidimensional fatigue-related measures utilized in the nonpharmacological treatment outcome literature, thus making comparisons between these studies untenable.
706
CANCER-RELATED FATIGUE
707
Table 1
Randomized Controlled Trials Sample, Design Characteristics, and Fatigue and Vigor/Vitality Outcomes
Variable
Total no. trials (published studies)
CRF aim/hypothesis
Yes
No
Design
CS
APb
FUc
Cancer type
BC
Mix
MM
Other
Stage
Early
Advanced
Mix
Mets.
NA
Treatment status
On
Off
Mix
NA
Includes POMS Fatigue and/or Vigor subscale
Resultsd
Positive
All
% with CRF aim/hypothesis
Negative
All
% with CRF aim/hypothesis
Equivalent
All
% with CRF aim/hypothesis
Psychosocial, includes
Exercise, includes
fatigue outcome:
fatigue outcome:
Systematic review n (%) Systematic review n (%)
50a (42)
17a (17)
Psychosocial, includes
vigor/vitality
Exercise, includes
outcome:Systematic
vigor/vitality outcome:
review n (%)
Systematic review n (%)
50a (39)
10a (9)
14 (28%)
36 (72%)
10 (59%)
7 (41%)
9 (18%)
41 (82%)
5 (50%)
5 (50%)
27 (54%)
23 (46%)
20 (40%)
7 (41%)
10 (59%)
33 (66%)
17 (34%)
21 (42%)
4 (40%)
6 (60%)
21 (42%)
19 (38%)
5 (10%)
5 (10%)
10 (59%)
3 (18%)
4 (23%)
28 (56%)
14 (28%)
4 (8%)
4 (8%)
7 (70%)
2 (20%)
0
1 (10%)
8 (16%)
3 (6%)
27 (54%)
7 (14%)
5 (10%)
6 (35%)
1 (6%)
8 (47%)
0
2 (12%)
13 (26%)
6 (12%)
20 (40%)
7 (14%)
4 (8%)
5 (50%)
0
3 (30%)
0
2 (20%)
25 (50%)
9 (18%)
7 (14%)
9 (18%)
33 (66%)
11 (65%)
5 (29%)
1 (6%)
0
6 (35%)
19 (38%)
15 (30%)
7 (14%)
9 (18%)
36 (72%)
6 (60%)
4 (40%)
0
0
6 (60%)
12 (24%)
7 (50%)
6 (35%)
3 (30%)
15 (30%)
3 (33.3%)
3 (30%)
2 (40%)
32 (64%)
5 (36%)
7 (41%)
4 (40%)
32 (64%)
4 (44.4%)
5 (50%)
2 (40%)
6 (12%)
2 (14%)
4 (24%)
3 (30%)
3 (6%)
2 (22.2%)
2 (20%)
1 (20%)
Note. CRF aim/hypothesis number of studies that include a cancer-related fatigue specific aim and/or hypothesis; Design: CS cross-sectional; AP
adjuvant prospective; FU follow-up; cancer type: BC breast cancer sample used; mix heterogeneous sample of cancer patients used; MM
malignant melanoma sample used; other includes samples based entirely on other types of cancers excluding mixed, BC, and MM samples; stage: mets.
metastatic disease; NA not available: study did not report disease staging of sample participants; treatment status: NA not available: study did not
report treatment status of sample participants. POMS Profile of Mood States.
a
Sample size characteristics and proportions based on total number of trials. b We chose to use the term adjuvant prospective (AP) design given that none
of the studies were strictly prospective, as baseline assessments occurred at varying intervals post cancer diagnosis and either prior to the commencement
of cancer treatment or soon after the start of treatment. Hence, study designs were categorized as AP if the intervention was administered at the same time
that participants commenced either their primary or adjuvant medical/cancer treatments (i.e., radiotherapy or chemotherapy). c Includes at least one longer
term follow-up in addition to the initial postintervention assessment published within the same article. d Results: As no exercise study reported on multiple
follow-up assessments postintervention, for comparative purposes, results for both the psychosocial and exercise studies are based on the assessment
findings conducted postintervention (i.e., excluding additional longer term follow up assessments); all inclusive of all studies with or without CRF
aim/hypothesis; % with CRF aim/hypothesis percentage of all studies that include CRF aim/hypothesis and that report positive, negative, and equivalent
findings, respectively.
708
Table 2
Fatigue and Vigor/Vitality Results According to Specific Therapeutic Components
Intervention
Psychosocial treatments: RCTs
Cognitivebehavioral therapy
Supportiveexpressive therapy
Counseling
Educational
Behavioral/relaxation
Massage
Restorative
Psychosocial treatments: Single-group design
Cognitivebehavioral therapy
Supportive counseling
Educational
Behavioral/relaxation
Massage/healing therapies
Restorative
Exercise treatments: RCTs
Multimodal
Walking
Bicycle
Cardiovascular/flexibility and/or strength
Resistance
Exercise treatments: Single-group design
Multimodal
Aerobic
% of studies reporting
positive fatigue results
% of studies reporting
positive vigor/vitality
results
19
7
9
7
5
1
2
26%
43%
11%
0%
0%
100%
100%
20
6
12
6
7
1
1
40%
17%
33%
33%
0%
0%
0%
9
6
0
3
2
4
67%
50%
0%
0%
50%
6
5
1
2
2
1
33%
60%
0%
0%
0%
100%
7
3
2
4
1
43%
33%
50%
0%
100%
5
3
1
1
0
40%
33%
0%
0%
7
4
43%
100%
5
1
40%
0%
particular, participants who received a combined group psychotherapy and home-based aerobic exercise intervention were found to have
a significant improvement in fatigue compared with participants who
only received the group psychotherapy intervention without the
home-based exercise therapy component (Courneya, Friedenreich,
Sela, Quinney, Rhodes, & Handman, 2003).
Only half of the non-RCT psychosocial group comparison studies included vigor/vitality as an outcome measure (Bailey, 1983;
Hernandez-Reif, Field, Ironson, Beutler, & Vera, 2005; Targ &
Levine, 2002). Two of these studies documented a significant
improvement in vigor for two of the interventions. Specifically,
Bailey (1983) found that a live music intervention significantly
improved vigor compared with taped music, whereas HernandezReif et al. (2005) reported that massage therapy significantly
increased vigor compared with participants who received progressive muscle relaxation training.
ment; however, only two of these trials were found to maintain the
reduction in fatigue at 3 months following completion of the intervention (Allison, Edgar, et al., 2004; Allison, Nicolau, Edgar, et al.,
2004; Cunningham, Lockwood, & Edmonds, 1993).
Seventeen psychosocial single-group design trials reported outcomes for vigor/vitality, and only one third of these interventions
included CRF as a specific aim and/or hypothesis of the study.
About one third of the single-group design trials (29%) reported
positive outcomes for vigor/vitality, and only one of these trials
included a specific CRF aim. Two of the six studies that reported
an improvement in vigor/vitality included an additional longer
term follow-up assessment; however, only one of these trials was
found to maintain improvements at 3 months following completion
of the intervention (Cunningham et al., 1993).
The single-group design psychosocial interventions were also classified according to their primary therapeutic approaches. The categories were the same as those used for the RCT psychosocial interventions. Two thirds of single-group designs that were CBT programs
(n 6) were found to significantly reduce fatigue. In addition, 50%
of supportive counseling (n 3) and restorative therapies (n 2)
were found to improve fatigue symptoms. However, none of the
behavioral/relaxation (n 3) and alternative (massage and body
healing; n 2) therapies were found to have significantly beneficial
effects on fatigue outcomes. Furthermore, the majority (60%; n 3)
of the supportive counseling and one third (n 2) of the CBT
intervention studies reported improved vigor/vitality. The one restorative therapeutic trial was also found to increase vigor/vitality, although none of the behavioral/relaxation (n 2), massage and body
CANCER-RELATED FATIGUE
709
Table 3
Single-Design Studies: Sample, Design Characteristics, and Fatigue and Vigor/Vitality Outcomes
Variable
Total no. trials (published studies)
CRF aim/hypothesis
Yes
No
Design
CS
APb
FUc
Cancer type
BC
Mix
MM
Other
Stage
Early
Advanced
Mix
Mets.
NA
Treatment status
On
Off
Mix
NA
Includes POMS Fatigue and/or
Vigor subscale
Resultsd
Positive
All
% with CRF aim/hypothesis
Negative
All
% with CRF aim/hypothesis
Psychosocial includes
fatigue outcome:
Systematic review
n (%)
24a (20)
10 (42%)
14 (58%)
7 (64%)
4 (36%)
16 (67%)
8 (33%)
5 (21%)
6 (35%)
4 (24%)
2 (33%)
4 (67%)
6 (54.5%)
5 (45.5%)
13 (76%)
4 (24%)
3 (18%)
4 (67%)
2 (33%)
8 (33%)
15 (63%)
0
1 (4%)
5 (45.5%)
5 (45.5%)
0
1 (9%)
4 (24%)
13 (76%)
0
0
1 (17%)
5 (83%)
0
0
3 (12%)
0
10 (42%)
0
11 (46%)
0
1 (9%)
7 (64%)
0
3 (27%)
2 (125)
0
7 (41%)
0
8 (47%)
0
0
4 (67%)
0
2 (33%)
10 (42%)
5 (21%)
3 (12%)
6 (25%)
14 (58%)
5 (45.5%)
3 (27.3%)
1 (9%)
2 (18.2%)
3 (27%)
6 (35%)
2 (12%)
2 (12%)
7 (41%)
13 (76%)
11 (46%)
4 (40%)
6 (54.5%)
6 (86%)
5 (29%)
1 (17%)
2 (33%)
1 (50%)
13 (54%)
6 (60%)
5 (45.5%)
1 (14%)
12 (71%)
5 (83%)
4 (67%)
1 (50%)
2 (33.3%)
1 (17.7)
1 (17.7)
2 (33.3)
3 (50%)
Note. CRF aim/hypothesis number of studies that included a cancer-related fatigue specific aim and/or hypothesis. POMS Profile of Mood States.
All inclusive of all studies with or without CRF aims/hypothesis; % with CRF aim/hypothesis percentage of studies that include CRF aim/hypothesis;
Design: CS cross-sectional; AP adjuvant prospective; FU includes at least one longer-term follow-up, in addition to initial post-intervention
assessment; Cancer Type: BC breast cancer sample used; Mix heterogeneous sample of cancer patients used; MM malignant melanoma sample
used; Other includes samples based entirely comprising of other types of cancers excluding mixed, BC, and MM samples; Stage: Mets metastatic
disease; NA not available: study did not report disease staging of sample participants; treatment status: NA not available: study did not report treatment
status of sample participants.
a
Sample size characteristics and proportions based on total number of trials. b We chose to use the term adjuvant prospective (AP) design given that none
of the studies were strictly prospective, as baseline assessments occurred at varying intervals post cancer diagnosis and either prior to the commencement
of cancer treatment or soon after the start of treatment. Hence, study designs were categorized as AP if the intervention was administered at the same time
that participants commenced either their primary or adjuvant medical/cancer treatments (i.e., radiotherapy or chemotherapy). c Includes at least one longer
term follow-up in addition to the initial postintervention assessment published within the same article. d Results: As no exercise study reported on multiple
follow-up assessments postintervention, for comparative purposes, results for both the psychosocial and exercise studies are based on the assessment
findings conducted postintervention (i.e., excluding additional longer term follow up assessments); all inclusive of all studies with or without CRF
aim/hypothesis; % with CRF aim/hypothesis percentage of all studies that include CRF aim/hypothesis and that report positive, negative, and equivalent
findings, respectively.
Exercise Interventions
Nineteen articles reported effectiveness of various physical exercise interventions with a fatigue-related outcome measure using
an RCT (n 17 studies) or non-RCT/CCT (n 2) group com-
710
Meta-Analytic Review
Methodological Quality of RCT Interventions
Standard criteria (Altman et al., 2001) regarding methodological
quality of the psychosocial and exercise RCT studies included in
the meta-analyses (including studies that reported fatigue and/or
vigor/vitality outcomes) were evaluated. The quality scores for the
psychosocial RCTs (n 43 publications) ranged from 1.5 to 7 out
of a maximum of 8 points. (As one might imagine, blinding of
participants to psychosocial or exercise interventions is almost
impossible, as participants must actively engage in these interventions; hence, no study received a perfect score of 8). All trials
specified participant inclusion criteria; however, 54% (n 24) did
not specify method of randomization used. In addition, for 74% of
the studies (n 32), participants were comparable at baseline or
adjustments were made to control for baseline differences; 63%
(n 27) included at least 80% of participants at the first follow-up
assessment following completion of the intervention; and 65%
(n 28) included multiple follow-up assessments. The main
limitations for the psychosocial RCTs were nonreporting or use of
blinding of outcome assessors (n 37; 86%); concealment of
CANCER-RELATED FATIGUE
711
Fatigue-Related Theory
Only a minority of psychosocial (15%) and exercise (19%)
studies specified a fatigue-related theory in substantiating the
intervention(s) evaluated to reduce fatigue. Psychosocial interventions that reported a specific fatigue-related theory were found to
have a significantly smaller effect size (.27) than exercise interventions that reported a fatigue-specific theory (.83, p .05).
Both psychosocial and exercise interventions that did not report a
fatigue-specific theory were found to have moderately small effect
sizes (.32). Psychosocial studies that did not report a fatiguespecific theory but included a CRF aim/hypothesis were found to
have a moderate effect size (.58) compared with the exercise
interventions that included a CRF aim/hypothesis (.15); this
interaction was significant ( p .01).
Table 4 presents a summary of the meta-analytic results comparing the effects of exercise interventions to psychosocial interventions on fatigue outcomes according to 14 main sets of poten-
Quality of study. Psychosocial interventions that were classified as having good methodological validity had a slightly larger
(text continues on page 716)
712
Table 4
Sample Size, Effect Size, Confidence Interval, and Significance for Fatigue Outcome for Each Study Variable
Effect size
(ES; weighted)
Variable
Overall (psychological and exercise studies)
Group comparison
Psychological
Exercise
CRF aim
Psychological
Exercise
Non-CRF aim
Psychological
Exercise
Fatigue-specific theory
Psychological (all)
Exercise (all)
No-fatigue specific theory
Psychological (all)
Exercise (all)
Fatigue-specific theory
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
No fatigue-specific theory
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
POMS Fatigue subscale used
Psychological (all)
Exercise (all)
Non-POMS fatigue measure used
Psychological (all)
Exercise (all)
POMS Fatigue subscale used
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Non-POMS fatigue measure used
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment modality: individual therapy
Psychological (all)
Exercise (all)
Treatment modality: group therapy
Psychological (all)
Exercise (all)
Treatment modality: individual therapy
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment modality: group therapy
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Baseline compatibility and/or adjustment made
Psychological (all)
Exercise (all)
95% CI
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
57 (N 4,621)
0.341
.000
0.402
0.280
41 (N 3,620)
16 (N 1,001)
0.313
0.415
.000
.000
0.377
0.599
0.249
0.231
13
9
0.484
0.375
.000
.005
0.571
0.637
0.398
0.114
28
7
0.234
0.465
.000
.000
0.318
0.711
0.150
0.219
6
3
0.268
0.826
.005
.000
0.454
1.100
0.081
0.551
35
13
0.321
0.319
.000
.001
0.395
0.510
0.247
0.128
4
2
3
0
0.292
0.220
0.826
.000
.700
.000
0.365
1.337
1.100
0.218
0.897
0.551
9
26
6
7
0.575
0.234
0.150
0.465
.000
.000
.346
.000
0.706
0.311
0.462
0.711
0.444
0.158
0.162
0.219
25
5
0.328
0.387
.000
.031
0.392
0.737
0.265
0.036
16
11
0.286
0.428
.000
.000
0.406
0.649
0.165
0.207
4
21
4
1
0.605
0.279
0.410
0.290
.000
.000
.047
.000
0.812
0.346
0.816
0.372
0.398
0.213
0.005
0.208
9
7
5
6
0.435
0.093
0.347
0.495
.000
.322
.064
.000
0.535
0.277
0.715
0.766
0.336
0.091
0.020
0.223
23
15
0.315
0.393
.000
.000
0.382
0.583
0.249
0.203
18
1
0.308
0.744
.000
.000
0.442
0.847
0.175
0.641
11
12
9
6
0.507
0.143
0.375
0.419
.000
.001
.005
.001
0.597
0.224
0.637
0.675
0.416
0.061
0.114
0.162
2
16
0
1
0.360
0.302
.317
.000
0.360
0.442
1.066
0.161
0.744
.000
0.847
0.641
29
12
0.285
0.307
.000
.004
0.341
0.515
0.229
0.100
Interaction:
p
.342
.102
.036
.002
.699
.588
.121
.355
.048
.147
CANCER-RELATED FATIGUE
713
Table 4 (continued )
Effect size
(ES; weighted)
Variable
No baseline compatibility and/or adjustment
made
Psychological (all)
Exercise (all)
Baseline compatibility and/or adjustment made
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
No baseline compatibility and/or adjustment
made
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
ITT and/or handling of missing data
Psychological (all)
Exercise (all)
No ITT or handling of missing data
Psychological (all)
Exercise (all)
ITT and/or handling of missing data
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
No ITT or handling of missing data
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
More than 80% retention at postassessment
Psychological (all)
Exercise (all)
Less than 80% retention at postassessment
Psychological (all)
Exercise (all)
More than 80% retention at postassessment
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Less than 80% retention at postassessment
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
AP design
Psychological (all)
Exercise (all)
CS design
Psychological (all)
Exercise (all)
AP design
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
CS design
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
95% CI
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
Interaction:
p
12
4
0.383
0.742
.001
.001
0.614
1.173
0.153
0.312
8
21
7
5
0.350
0.261
0.320
0.289
.000
.000
.032
.030
0.444
0.332
0.613
0.551
0.256
0.189
0.027
0.028
5
7
2
2
0.721
0.149
0.570
0.914
.000
.289
.164
.000
0.797
0.425
1.374
1.253
0.645
0.127
0.233
0.575
11
9
0.523
0.154
.000
.194
0.638
0.386
0.407
0.078
30
7
0.235
0.753
.000
.000
0.305
0.989
0.165
0.518
5
6
6
3
0.557
0.494
0.192
0.065
.000
.000
.239
.203
0.688
0.662
0.511
0.164
0.426
0.325
0.127
0.035
8
22
3
4
0.431
0.163
0.743
0.758
.000
.001
.001
.000
0.512
0.255
1.194
0.933
0.350
0.070
0.292
0.583
28
12
0.340
0.345
.000
.001
0.424
0.554
0.256
0.136
13
4
0.255
0.629
.000
.008
0.379
1.090
0.131
0.168
8
20
7
5
0.512
0.271
0.320
0.379
.000
.000
.032
.008
0.622
0.367
0.613
0.658
0.402
0.174
0.027
0.100
5
8
2
2
0.445
0.141
0.570
0.688
.001
.139
.164
.085
0.701
0.329
1.374
1.472
0.188
0.046
0.233
0.096
.331
15
9
0.223
0.546
.000
.000
0.332
0.726
0.114
0.367
.031
26
7
0.366
0.243
.000
.084
0.453
0.520
0.278
0.033
5
10
7
2
0.422
0.123
0.447
0.904
.000
.134
.000
.000
0.569
0.284
0.647
1.256
0.276
0.038
0.247
0.551
8
18
2
5
0.528
0.295
0.125
0.283
.000
.000
.784
.000
0.709
0.398
1.017
0.442
0.347
0.192
0.767
0.124
.804
.036
.000
.854
.257
.124
.221
.008
.237
(table continues)
714
Table 4 (continued )
Effect size
(ES; weighted)
Variable
Type of cancer: 100% breast cancer sample
Psychological (all)
Exercise (all)
Type of cancer: mixed or non-BC sample
Psychological (all)
Exercise (all)
Type of cancer: 100% breast cancer sample
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Type of cancer
Mixed or non-BC sample
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
On treatment at initial assessment
Psychological (all)
Exercise (all)
Not on treatment at initial assessment
Psychological (all)
Exercise (all)
On treatment at initial assessment
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (Non-CRF aim/hypothesis)
Not on treatment at initial assessment
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Good validitya
Psychological (all)
Exercise (all)
Poor validityb
Psychological (all)
Exercise (all)
Good validitya
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Poor validityb
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Setting: institution
Psychological (all)
Exercise (all)
Setting: home or combination
Psychological (all)
Exercise (all)
Setting: institution
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
95% CI
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
Interaction:
p
18
9
0.281
0.618
.000
.000
0.391
0.850
0.171
0.386
.010
23
7
0.338
0.155
.000
.330
0.421
0.467
0.256
0.157
5
13
4
5
0.619
0.155
0.679
0.570
.000
.007
.000
.008
0.833
0.267
1.000
0.992
0.404
0.042
0.359
0.147
8
15
5
2
0.406
0.303
0.132
0.212
.000
.000
.531
.440
0.506
0.429
0.545
0.751
0.306
0.176
0.281
0.327
.577
26
10
0.243
0.566
.000
.000
0.327
0.736
0.159
0.396
.003
15
6
0.437
0.161
.000
.282
0.542
0.454
0.332
0.132
9
17
7
3
0.477
0.118
0.447
0.844
.000
.033
.000
.000
0.575
0.226
0.647
1.025
0.379
0.009
0.247
0.664
4
11
2
4
0.509
0.413
0.125
0.164
.015
.000
.784
.008
0.917
0.531
1.017
0.286
0.100
0.295
0.767
0.042
23
8
0.355
0.275
.000
.036
0.432
0.533
0.277
0.018
18
8
0.260
0.557
.000
.001
0.383
0.875
0.137
0.238
6
17
6
2
0.513
0.298
0.376
0.027
.000
.000
.022
.400
0.658
0.382
0.699
0.036
0.368
0.214
0.054
0.091
7
11
3
5
0.461
0.132
0.373
0.665
.000
.143
.258
.000
0.592
0.310
1.020
0.888
0.330
0.045
0.274
0.442
31
6
0.345
0.336
0.000
0.016
0.424
0.609
0.266
0.063
10
10
0.216
0.462
0.001
0.001
0.345
0.731
0.086
0.194
11
20
2
4
0.490
0.265
0.250
0.380
.000
.000
0.354
0.032
0.585
0.365
0.779
0.726
0.396
0.165
0.279
0.034
.174
.001
.692
.071
.531
.041
.259
.225
CANCER-RELATED FATIGUE
715
Table 4 (continued )
Effect size
(ES; weighted)
Variable
Setting: home or combination
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment duration: 6 weeks
Psychological (all)
Exercise (all)
Treatment duration: 6 to 8 weeks
Psychological (all)
Exercise (all)
Treatment duration: 8 weeks
Psychological (all)
Exercise (all)
Treatment duration: 6 weeks
Psychological (CRF aim/hypothesis)
Psychological (non-CRF)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment duration: 6 to 8 weeks
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment duration: 8 weeks
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Specific treatment approaches: psychosocial (all)
CBT
Supportive expressive
Behavioral/ relaxation
Counseling
Educational
Massage
Restorative
Specific treatment approaches: psychosocial
(CRF aim/hypothesis)
CBT
Supportive expressive
Behavioral/ relaxation
Counseling
Educational
Massage
Restorative
Specific treatment approaches: psychosocial
(non-CRF aim)
CBT
Supportive expressive
Behavioral/ relaxation
Counseling
Educational
Massage
Restorative
Specific treatment approaches: exercise (all)
Walking only
Multimodal
Bicycle
Cardiovascular and/or flexibility
Resistance
95% CI
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
0.455
0.075
0.945
0.045
0.411
0.019
0.745
0.068
12
0
0.402
0.000
0.560
0.245
21
3
0.264
0.346
0.000
0.080
0.357
0.735
0.170
0.042
8
13
0.313
0.431
0.000
0.000
0.458
0.644
0.168
0.218
7
5
0
0
0.488
0.272
.000
.243
0.601
0.739
0.375
0.187
3
18
2
1
0.321
0.254
0.155
0.730
.062
.000
.376
.000
0.657
0.357
0.498
0.773
0.016
0.151
0.188
0.687
3
5
7
6
0.638
0.103
0.638
0.413
.000
.000
.000
.000
0.778
0.158
0.778
0.596
0.497
0.048
0.497
0.231
17
6
4
7
4
1
2
0.304
0.445
0.199
0.294
0.110
0.680
0.524
.000
.000
.013
.003
.124
.000
.107
0.447
0.638
0.355
0.487
0.250
0.739
1.161
0.161
0.252
0.043
0.100
0.030
0.621
0.113
3
2
2
2
1
1
2
0.431
0.617
0.328
0.540
0.250
0.680
0.524
.000
.000
.320
.002
.000
.000
.107
0.512
0.813
0.975
0.873
0.261
0.739
1.161
0.350
0.422
0.319
0.206
0.239
0.621
0.113
14
4
2
5
3
0
0
0.270
0.354
0.095
0.193
0.063
.001
.000
.317
.015
.274
0.432
0.504
0.281
0.349
0.176
0.107
0.204
0.091
0.038
0.050
3
6
2
4
1
0.496
0.627
.004
0.219
0.520
.006
.000
.757
.262
.000
0.849
0.785
0.020
0.601
0.533
0.143
0.469
0.027
0.163
0.507
Interaction:
p
.301
(table continues)
716
Table 4 (continued )
Effect size
(ES; weighted)
Variable
Specific treatment approaches: exercise
(CRF aim/hypothesis)
Walking only
Multimodal
Bicycle
Cardiovascular and/or flexibility
Resistance
Specific treatment approaches: Exercise
(non-CRF aim/hypothesis)
Walking only
Multimodal
Bicycle
Cardiovascular and/or flexibility
Resistance
95% CI
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
3
2
1
2
1
0.496
0.372
0.020
0.325
0.520
.006
.077
.238
.620
.000
0.849
0.783
0.013
1.608
0.533
0.143
0.040
0.053
0.959
0.507
0
4
1
2
0
0.758
0.005
0.113
.000
.610
.520
0.933
0.024
0.456
0.583
0.014
0.230
Interaction:
p
Note. CRF cancer-related fatigue; POMS Profile of Mood States: ITT intention to treat; AP adjuvant prospective; CS cross-sectional.
Good validity is represented by a score above 4 points (out of 8). b Poor validity is represented by a score of less than 4.5 points.
p .05. p .01.
a smaller effect size (.57) than exercise studies that did not
include a CRF aim/hypothesis (.91).
ITT analyses and handling of missing data. Psychosocial interventions that reported use of ITT analyses and/or reported how
participant attrition or missing data were handled were found to
have a moderately stronger effect size (.52) than the psychosocial interventions that did not (.24). The opposite pattern
emerged for exercise interventions, where a stronger effect was
noted for interventions that did not report the use of ITT analyses
and/or handling of missing data (.75) compared with exercise
studies that did (.15). This group interaction effect was found to
be significant ( p .001), but no significant interaction effect was
evident between studies use of ITT analyses, reporting of handling missing data, and dropouts and CRF aim/hypothesis status.
Participant retention rates at the initial assessment following
completion of intervention. Psychosocial interventions that included postassessment data for at least 80% of participants who
were recruited and assessed at both baseline and postassessment
were found to have a stronger (although moderately small) effect
size (.34) than the psychosocial interventions in which less than
80% of participants completed the post-assessment (.26). Once
again, the opposite pattern was evident for exercise interventions;
a stronger effect was noted for interventions where fewer than 80%
of participants completed the postassessment (.63) compared
with exercise studies with data for more than 80% of participants
at postassessment (.35). However, this group interaction effect
was found to be nonsignificant. Similarly, no significant interaction was evident between intervention category, participant retention rates, and whether studies included a CRF aim/hypothesis.
Study Design
Psychosocial interventions that used a cross-sectional (CS) design were found to have a moderately stronger effect size (.37)
than the psychosocial interventions that used an adjuvant prospective (AP) design (.22). In contrast, a stronger effect size was
CANCER-RELATED FATIGUE
Participant Characteristics
Cancer type. Psychosocial interventions that solely included
breast cancer patients were found to have a smaller effect size
(.28) than psychosocial interventions that included mixed samples (including breast cancer patients) or focused explicitly on
non breast cancer samples (e.g., prostate or head and neck malignancies; .34). Conversely, exercise interventions that explicitly
included only breast cancer patients were found to have a stronger
effect size (.62) than exercise studies that included mixed samples (including breast cancer patients) or explicitly focused on
non breast cancer samples (.16). The group interaction effect
was found to be significant ( p .05), although no significant
interaction effect emerged between intervention category, cancer
type, and whether studies included a CRF aim/hypothesis.
Medical/cancer treatment status. Psychosocial studies that explicitly included participants who were receiving primary or adjuvant medical treatment (e.g., radiation or chemotherapy) at the
time of the psychosocial intervention were found to have a smaller
effect size (.24) than psychosocial interventions that included
participants who were not on primary or adjuvant treatment at the
time of the intervention or included mixed samples of participants
(of whom some were on treatment and others had completed
treatment; .44). Once again, the opposite pattern emerged for
exercise interventions, where a moderate effect size was revealed
for studies that explicitly included participants receiving primary
or adjuvant medical treatment at the time the exercise program was
administered (.57) compared with exercise studies that included
participants who were not on primary or adjuvant treatment at the
time of the intervention or included mixed samples of participants
(of whom some were on treatment and others had completed
treatment) at the time of the intervention (.16). The group
interaction effect was found to be significant ( p .05). In addition, a significant interaction effect was found between studies that
included participants who were receiving primary or adjuvant
treatment at the time of the intervention and whether they included
a CRF aim/hypothesis ( p .01). In particular, psychosocial studies that included participants who were receiving primary or adjuvant treatment at the time of the intervention and also included
a CRF aim/hypothesis were found to have moderately larger effect
size (.48) than psychosocial interventions that did not include a
CRF aim/hypothesis (.12). In contrast, exercise studies that
included participants who were receiving primary or adjuvant
717
Intervention Variables
Treatment modality. The effect size for psychosocial interventions that were administered on an individual basis (.32) was
comparable to the effect size for interventions administered in a
group format (.31). Only one exercise study included a group
exercise format, and a relatively strong effect size (.74) was
found compared with individually administered exercise programs,
which were found to have a relatively moderate effect size (.39).
Moreover, a significant interaction effect was found between studies that relied on individually administered interventions and a
CRF aim/hypothesis status ( p .05). Specifically, psychosocial
studies that included a CRF aim/hypothesis were found to have a
moderately larger effect size (.51) than psychosocial studies that
did not include a CRF aim/hypothesis (.14). Conversely, exercise studies that did not include a CRF aim/hypothesis were found
to have a slightly larger, moderate effect size (.42) than exercise
interventions that did include a CRF aim/hypothesis (.38).
Treatment setting. Psychosocial interventions that were administered in a hospital setting were found to have a slightly larger
moderate effect size (.35) than psychosocial interventions that
were administered in the home or were administered in combined
settings (i.e., some sessions were conducted at the hospital while
others were conducted at home; .22). Exercise interventions that
were conducted at home or were administered in combined settings were found to have a moderate effect size (.46) compared
with the smaller effect size found for exercise programs that were
explicitly administered in a hospital setting (.34). This group
interaction effect was not significant. Similarly, no significant
interaction effect was found between intervention category, setting
location, and CRF aim/hypothesis status.
Treatment duration. Psychosocial interventions that consisted
of fewer than six sessions were found to have a moderate effect
size (.40) compared with the smaller effect for psychosocial
interventions that included between 6 and 8 sessions (.26) as
well as those with more than 8 sessions (.31). None of the
exercise interventions consisted of fewer than 6 sessions. Exercise
programs that consisted of more than 8 sessions were found to
have a moderate effect size (.43) compared with the smaller
effect size for interventions that consisted of 6 to 8 sessions
(.35).
718
psychosocial and exercise interventions on vigor/vitality outcomes. The overall effect sizes for the 36 psychosocial intervention trials (N 3,460) ranged from .19 to 1.45 (where positive
indices indicated higher vigor/vitality symptoms postintervention), with a weighted pooled mean effect size of .37. This
effect was significant (z 10.49; p .001) and clinically meaningful. Thus, the overall effect of psychosocial interventions on
vigor/vitality was small to moderate. The overall effect sizes for
the seven exercise intervention trials (N 395) ranged from .27 to
1.42, with a weighted pooled mean effect size of .69. This effect
was also significant (z 5.28, p .001) and clinically meaningful. Hence, the overall effect of physical exercise interventions
on vigor/vitality was in the moderate to strong range. In comparing
the difference in effects on vigor/vitality for exercise interventions
with psychosocial interventions, we found a significant group
difference showing that exercise interventions were stronger than
psychosocial interventions ( p .05).
Possible publication bias was also evaluated; no evidence of
bias was found for either the psychosocial, t(35) 1.07, p .05,
or the exercise, t(6) 0.79, p .05 trials. These results provide
no support for the possibility that the significant effect sizes
reported for vigor/vitality outcomes were due to publication bias.
Vigor/Vitality-Related Theory
Only a minority of psychosocial (8%) and exercise (29%) studies specified a vigor/vitalityrelated theory in substantiating the
intervention(s) evaluated to enhance vigor/vitality. Psychosocial
(text continues on page 726)
CANCER-RELATED FATIGUE
719
Table 5
Psychological Interventions: Sample Size, Effect Size, Confidence Interval, and Significance for Fatigue Outcome for Each Study
Variable at Longer Term Follow-ups
Effect size
(ES; weighted)
Variable
Psychological studies with additional follow-upsa
Within 4 months
6 months
6 months
CRF aim
Within 4 months
6 months
6 months
Non-CRF aim
Within 4 months
6 months
6 months
Fatigue-specific theory
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
No fatigue-specific theory
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
POMS Fatigue subscale
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Non-POMS fatigue measure
Within 4 months, with CRF Aim
within 4 months, with non-CRF aim
Treatment modality: individual therapy
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment modality: group therapy
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Baseline compatibility and/or adjustment made
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
No baseline compatibility and/or adjustment
made
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
ITT and/or handling of missing data
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
95% CI
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
13
7
5
1
0.421
0.514
0.304
0.360
.000
0.006
0.000
0.000
0.600
0.881
0.464
0.377
0.242
0.147
0.144
0.343
2
0
0
0.785
0.500
3.068
1.498
5
5
1
0.405
0.304
0.360
0.000
0.000
0.000
0.618
0.464
0.377
0.193
0.144
0.343
1
1
0.380
1.050
0.000
0.000
0.364
1.093
0.396
1.007
1
4
0
5
0
1
1.950
0.245
0.000
0.001
1.991
0.383
1.909
0.106
0.304
0.000
0.464
0.144
0.360
0.000
0.377
0.343
0.290
0.000
0.431
0.149
0.304
0.000
0.464
0.144
0.360
0.000
0.377
0.343
2
2
0.785
0.580
0.500
0.217
3.068
1.501
1.498
0.341
1
2
0
1
0
1
1.950
0.358
0.000
0.000
1.991
0.374
1.909
0.343
0.360
0.000
0.377
0.343
0.360
0.000
0.377
0.343
1
3
0
4
0
0
0.380
0.439
0.000
0.018
0.364
0.803
0.396
0.075
0.290
0.003
0.481
0.098
1
4
0
5
0
1
1.950
0.245
0.000
0.001
1.991
0.383
1.909
0.106
0.304
0.000
0.464
0.144
0.360
0.000
0.377
0.343
1
1
0.380
1.050
0.000
0.000
0.364
1.093
0.396
1.007
1
1
1.950
0.360
0.000
0.000
1.991
0.377
1.909
0.343
(table continues)
0
3
0
5
0
1
720
Table 5 (continued )
Effect size
(ES; weighted)
Variable
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
No ITT and/or handling of missing data
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
More than 80% retention at postassessment
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Less than 80% retention at postassessment
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
AP design
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
CS design
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
100% breast cancer sample
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Mixed or non-BC sample:
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
On treatment at initial assessment
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
No. of study
trials: K
95% CI
(random effects)
ES
p (random)
Lower
Upper
0.475
0.000
0.700
0.249
0.360
0.000
0.377
0.343
1
4
0
3
0
0
0.380
0.417
0.000
0.005
0.364
0.706
0.396
0.128
0.188
0.000
0.270
0.105
1
2
0
4
0
1
1.950
0.235
0.000
0.060
1.991
0.480
1.909
0.010
0.340
0.001
0.548
0.131
0.360
0.000
0.377
0.343
1
3
0
1
0
0
0.380
0.520
0.000
0.038
0.364
1.011
0.396
0.028
0.160
0.000
0.181
0.139
0.360
0.000
0.377
0.343
0.360
0.000
0.377
0.343
0.360
0.000
0.377
0.343
2
4
0
4
0
0
0.785
0.417
0.500
0.005
3.068
0.706
1.498
0.128
0.290
0.003
0.481
0.098
2
3
0
2
0
0
0.785
0.520
0.500
0.038
3.068
1.011
1.498
0.028
0.229
0.001
0.366
0.092
0.235
0.060
0.480
0.010
0.353
0.006
0.607
0.099
0.360
0.000
0.377
0.343
0.785
0.479
0.500
0.001
3.068
0.755
1.498
0.204
0.260
0.009
0.456
0.064
0.360
0.000
0.377
0.343
0
2
0
1
0
1
0
1
0
1
0
2
0
3
0
1
2
4
0
2
0
1
0
CANCER-RELATED FATIGUE
721
Table 5 (continued )
Effect size
(ES; weighted)
Variable
Not on treatment at initial assessment
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Good validity scoresb
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Poor validityc
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Institutional setting
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Home or combined home and institution setting
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment duration: Less than 6 sessions
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment duration: 6 to 8 sessions
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment duration: 8 sessions
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
Specific treatment approaches: within 4 months
CBT
Counseling
Specific treatment approaches: 6 months
CBT
Specific treatment approaches:
6 months
CBT
No. of study
trials: K
95% CI
(random effects)
ES
p (random)
Lower
Upper
0.110
0.000
0.120
0.100
0.333
0.040
0.651
0.015
1
2
4
0
1
1.950
0.235
0.340
0.000
0.060
0.001
1.991
0.480
0.548
1.909
0.010
0.131
0.360
0.000
0.377
0.343
1
3
0
1
0.380
0.520
0.000
0.038
0.364
1.011
0.396
0.028
0.160
0.000
0.181
0.139
0.210
0.010
0.369
0.051
0.304
0.000
0.464
0.144
0.360
0.000
0.377
0.343
0.785
0.700
0.500
0.046
3.068
1.386
1.498
0.014
0.360
0.000
0.377
0.343
0.360
0.000
0.377
0.343
0.360
0.000
0.377
0.343
0.785
0.605
0.500
0.174
3.068
1.477
1.498
0.267
0.290
0.003
0.481
0.098
0
2
0.229
0.056
0.465
0.006
6
1
0.275
1.950
0.051
0.000
0.550
1.991
0.001
1.909
0.304
0.000
0.464
0.144
0.360
0.000
0.377
0.343
0
1
0
3
0
0
0
3
0
5
0
1
2
2
0
0
0
0
0
1
0
1
0
1
2
2
0
4
0
0
Note. CRF cancer-related fatigue; POMS Profile of Mood States; CBT cognitive behavioral therapy.
The effects from the psychological studies which included at least one additional following assessment were categorized according to short-term (i.e.,
follow-up within 4 months), medium-range (6-month follow-up), and longer term (follow-up of more than 6 months) effects. b Good validity score
above 4 points (out of 8). c Poor validity less than 4.5 points.
p .05. p .01.
a
722
Table 6
Sample Size, Effect Size, Confidence Interval, and Significance for Vigor/Vitality Outcome for Each Study Variable
Effect size (ES)
(weighted)
Variable
Overall (psychological and exercise studies)
Group comparison
Psychological
Exercise
CRF aim
Psychological
Exercise
Non-CRF aim
Psychological
Exercise
Vigor/vitality specific theory
Psychological (all)
Exercise (all)
No vigor/vitality specific theory
Psychological (all)
Exercise (all)
Vigor/vitality specific theory
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
No vigor/vitality specific theory
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
POMS Vigor subscale used
Psychological (all)
Exercise (all)
Non-POMS vigor/vitality measure used
Psychological (all)
Exercise (all)
POMS Vigor subscale used
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Non-POMS vigor/vitality measure used
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment modality: individual therapy
Psychological (all)
Exercise (all)
Treatment modality:
group therapy
Psychological (All)
Exercise (All)
Treatment modality:
individual therapy
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment modality:
group therapy
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
95% confidence
interval
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
43 (N 3,855)
0.421
.000
0.354
0.489
36 (N 3,460)
7 (N 395)
0.369
0.692
.000
.000
0.300
0.435
0.438
0.949
8
3
0.502
0.790
.000
.000
0.391
0.443
0.614
1.137
28
4
0.332
0.613
.000
.000
0.243
0.385
0.421
0.840
3
2
0.290
0.974
.000
.000
0.134
0.788
0.446
1.160
33
5
0.377
0.570
.000
.000
0.295
0.399
0.460
0.742
3
0
2
0
0.290
.000
0.134
0.446
0.974
.000
0.788
1.160
5
28
1
4
0.640
0.332
0.420
0.613
.000
.000
.000
.000
0.437
0.243
0.387
0.385
0.842
0.421
0.453
0.840
27
3
0.387
0.660
.000
.005
0.265
0.202
0.508
1.119
9
4
0.317
0.716
.000
.000
0.252
0.355
0.382
1.077
4
23
2
1
0.585
0.353
0.745
0.490
.000
.000
.022
.000
0.400
0.220
0.108
0.408
0.771
0.485
1.382
0.572
4
5
1
3
0.425
0.238
0.880
0.654
.000
.000
.000
.000
0.277
0.172
0.857
0.382
0.573
0.304
0.903
0.926
22
7
0.294
0.692
.000
.000
0.225
0.435
0.364
0.949
14
0
0.484
.000
0.355
0.613
7
15
3
4
0.470
0.214
0.790
0.613
.000
.000
.000
.000
0.355
0.121
0.443
0.385
0.585
0.308
1.137
0.840
1
13
0
0
0.720
0.466
.000
.000
0.655
0.332
0.785
0.600
Interaction:
p
.023
.991
.169
.231
.690
.966
.952
.748
CANCER-RELATED FATIGUE
723
Table 6 (continued )
Effect size (ES)
(weighted)
Variable
Baseline compatibility and/or adjustment made
Psychological (all)
Exercise (all)
No baseline compatibility and/or adjustment made
Psychological (all)
Exercise (all)
Baseline compatibility and/or adjustment made
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
No baseline compatibility and/or adjustment made
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
ITT and/or handling of missing data
Psychological (all)
Exercise (all)
No ITT or handling of missing data
Psychological (all)
Exercise (all)
ITT and/or handling of missing data
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
No ITT or handling of missing data
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
More than 80% retention at postassessment
Psychological (all)
Exercise (all)
Less than 80% retention at postassessment
Psychological (all)
Exercise (all)
More than 80% retention at postassessment
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Less than 80% retention at postassessment
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
AP design
Psychological (all)
Exercise (all)
CS design
Psychological (all)
Exercise (all)
AP design
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
CS design
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
95% confidence
interval
(random effects)
No. of study
trials: K
ES
p (random)
Lower
Upper
27
7
0.340
0.692
.000
.000
0.246
0.435
0.433
0.949
9
0
0.453
.000
0.373
0.532
4
23
3
4
0.320
0.343
0.790
0.613
.000
.000
.000
.000
0.184
0.216
0.443
0.385
0.455
0.471
1.137
0.840
4
5
0
0
0.707
0.279
.000
.000
0.594
0.199
0.819
0.360
12
5
0.513
0.474
.000
.000
0.412
0.210
0.614
0.738
24
2
0.297
1.241
.000
.000
0.198
0.898
0.397
1.584
4
8
2
3
0.425
0.557
0.650
0.341
.000
.000
.005
.000
0.277
0.401
0.199
0.270
0.573
0.713
1.101
0.412
4
20
1
1
0.585
0.241
1.070
1.420
.000
.000
.000
.000
0.400
0.132
1.025
1.311
0.771
0.350
1.115
1.529
21
6
0.387
0.726
.000
.000
0.239
0.443
0.535
1.008
15
1
0.333
0.490
.000
.000
0.285
0.408
0.382
0.572
3
18
3
3
0.557
0.358
0.790
0.654
.000
0.000
.000
.000
0.377
0.193
0.443
0.382
0.736
0.523
1.137
0.926
5
10
0
1
0.467
0.278
.000
.000
0.334
0.231
0.599
0.326
0.490
.000
0.408
0.572
14
4
0.219
0.517
.000
.000
0.120
0.242
0.318
0.791
22
3
0.466
0.928
.000
.000
0.376
0.551
0.556
1.304
1
13
2
2
0.420
0.203
0.745
0.290
.000
.000
.022
.000
0.397
0.102
0.108
0.251
0.443
0.304
1.382
0.329
7
15
0.515
0.443
.000
.000
0.394
0.320
0.635
0.565
Interaction:
p
.578
.000
.285
.065
.619
.866
.543
.562
.744
(table continues)
724
Table 6 (continued )
No. of study
trials: K
ES
p (random)
1
2
0.880
0.954
20
5
95% confidence
interval
(random effects)
Lower
Upper
.000
.040
0.857
1.866
0.903
2.052
0.224
0.604
.000
.000
0.144
0.285
0.303
0.923
16
2
0.549
0.919
.000
.066
0.443
0.061
0.656
1.899
5
15
2
3
0.467
0.145
0.974
0.341
.000
.006
.000
.000
0.334
0.042
0.788
0.270
0.599
0.248
1.160
0.412
3
13
1
1
0.557
0.547
0.420
1.420
.000
.000
.000
.000
0.377
0.427
0.387
1.311
0.736
0.668
0.453
1.529
20
4
0.272
0.517
.000
.000
0.194
0.242
0.350
0.791
16
3
0.492
0.928
.000
.000
0.372
0.551
0.612
1.304
3
17
2
2
0.557
0.222
0.745
0.290
.000
.000
.022
.000
0.377
0.140
0.108
0.251
0.736
0.304
1.382
0.329
5
11
1
2
0.467
0.500
0.880
0.954
.000
.000
.000
.040
0.334
0.319
0.857
0.043
0.599
0.681
0.903
1.866
23
4
0.408
0.632
.000
.001
0.317
0.268
0.498
0.996
13
3
0.299
0.774
.000
.002
0.201
0.273
0.397
1.275
5
18
2
2
0.422
0.403
0.974
0.290
.000
.000
.000
.000
0.290
0.280
0.788
0.251
0.554
0.527
1.160
0.329
3
10
1
2
0.661
0.201
0.420
0.954
.000
.000
.000
.040
0.580
0.095
0.387
0.043
0.743
0.308
0.453
1.866
26
4
0.444
0.653
.000
.000
0.341
0.398
0.548
0.907
10
3
0.180
0.740
.000
.002
0.093
0.264
0.267
1.216
5
21
1
3
0.640
0.399
0.420
0.738
.000
.000
.000
.009
0.437
0.283
0.387
0.185
0.842
0.514
0.453
1.290
Interaction:
p
.959
.159
.05
.479
.693
.925
.401
.088
.007
.217
.220
CANCER-RELATED FATIGUE
725
Table 6 (continued )
Effect size (ES)
(weighted)
Variable
Setting: home or combination
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment duration: 6 weeks
Psychological (all)
Exercise (all)
Treatment duration: 6 to 8 weeks
Psychological (all)
Exercise (all)
Treatment duration: 8 weeks
Psychological (all)
Exercise (all)
Treatment duration: 6 weeks
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment duration: 6 to 8 weeks
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Treatment duration: 8 weeks
Psychological (CRF aim/hypothesis)
Psychological (non-CRF aim/hypothesis)
Exercise (CRF aim/hypothesis)
Exercise (non-CRF aim/hypothesis)
Specific treatment approaches: psychosocial (all)
CBT
Supportive expressive
Behavioral/ relaxation
Counseling
Educational
Massage
Restorative
Specific treatment approaches: psychosocial
(CRF aim/hypothesis)
CBT
Supportive expressive
Behavioral/ relaxation
Counseling
Educational
Massage
Restorative
Specific treatment approaches: psychosocial
(non-CRF aim)
CBT
Supportive expressive
Behavioral/ relaxation
Counseling
Educational
Massage
Restorative
Specific treatment approaches: exercise (all)
Walking only
Multimodal
Bicycle
Cardiovascular and/or flexibility
Resistance
No. of study
trials: K
ES
p (random)
95% confidence
interval
(random effects)
Lower
Upper
3
7
2
1
0.290
0.133
0.974
0.270
.000
.026
.000
.000
0.134
0.016
0.788
0.246
0.446
0.249
1.160
0.249
9
0
0.473
.000
0.367
0.579
15
2
0.362
0.745
.000
.022
0.171
0.108
0.552
1.382
12
5
0.296
0.671
.000
.000
0.237
0.356
0.355
0.986
6
3
0
0
0.479
0.460
.000
.002
0.355
0.162
0.604
0.757
0
15
2
0
0.362
0.745
.000
.022
0.171
0.108
0.552
1.382
2
10
1
4
0.568
0.242
0.880
0.613
.000
.000
.000
.000
0.274
0.185
0.857
0.385
0.862
0.299
0.903
0.840
14
5
4
8
3
1
1
0.472
0.377
0.136
0.317
0.183
0.330
0.850
.000
.004
.220
.000
.107
.000
.000
0.362
0.120
0.082
0.171
0.039
0.272
0.785
0.583
0.634
0.355
0.462
0.406
0.388
0.915
2
1
1
2
1
0
1
0.501
0.720
0.610
0.275
0.320
.000
.000
.000
.058
.000
0.306
0.655
0.488
0.009
0.313
0.696
0.785
0.732
0.559
0.327
0.850
.000
0.785
0.915
12
4
3
6
2
1
0
0.467
0.292
0.010
0.332
0.115
0.330
.000
.036
.934
.009
.555
.000
0.347
0.019
0.247
0.084
0.267
0.272
0.587
0.565
0.227
0.581
0.495
0.388
3
4
0
0
0
0.549
0.800
.004
.000
0.171
0.469
0.928
1.131
Interaction:
p
.014
(table continues)
726
Table 6 (continued )
No. of study
trials: K
95% confidence
interval
(random effects)
ES
p (random)
Lower
Upper
1
2
0
0
0
1.070
0.650
.000
.005
1.025
0.199
1.115
1.101
2
2
0
0
0
0.290
0.954
.000
.040
0.274
0.043
0.307
1.866
Interaction:
p
Note. CRF cancer-related fatigue; POMS Profile of Mood States; ITT intention to treat; AP adjuvant prospective: CS cross-sectional; CBT
cognitive behavioral therapy.
Good validity is represented by a score above 4 points (out of 8). b Poor validity is represented by a score of less than 4.5 points.
hypothesis (.42). The reverse pattern was found for exercise interventions. That is, exercise studies with poor methodological quality that did not specify a CRF aim/hypothesis were found to have
a much larger effect size (.95) than exercise studies that included
a CRF aim/hypothesis (.20). Although no significant interaction
effect was found between good methodological quality and CRF
aim/hypothesis status ( p .08), exercise studies with good methodological quality that did not specify a CRF aim/hypothesis were
found to have a much larger effect size (.97) than those that did
stipulate a CRF aim/hypothesis (.29). Psychosocial studies with
good methodological quality were found to have a moderate effect
size regardless of whether they did (.42) or did not (.40) specify a
CRF aim/hypothesis.
Baseline compatibility. Psychosocial intervention studies that
did not specify whether participants allocated to the two groups
(active intervention vs. control arm) were comparable at baseline
assessment on demographic and outcome variables, or did not
report whether adjustments were made in the analyses when baseline differences emerged, were found to have a somewhat larger
effect size (.45) than those that did (.34). In contrast, all of the
exercise interventions that reported a vigor/vitality outcome documented that participants were comparable at baseline or made
appropriate adjustments for group differences on demographic
and/or baseline variables.
ITT analyses and handling of missing data. The effect size for
psychosocial interventions that reported the use of ITT analyses
and/or reported participant attrition or handling of missing data
was moderately larger (.51) than the psychosocial interventions
that did not (.30). The two exercise studies that did not report the
use of ITT analyses and/or handling of missing data were found to
have very strong effect sizes (1.24) compared with the exercise
interventions that did (.47). The interaction of this variable with
intervention category was significant ( p .001). However, no
CANCER-RELATED FATIGUE
significant interaction effect was found between intervention category, ITT, and CRF aim/hypothesis status.
Participant retention rates at the initial assessment following
completion of intervention. Psychosocial intervention studies
that included postassessment data for at least 80% of participants
were found to have an effect size (.39) comparable to that of
psychosocial interventions in which fewer than 80% of participants completed the postassessment (.34). For exercise interventions that included postassessment data for at least 80% of participants, a stronger effect size (.73) was found compared with the
only exercise study in which fewer than 80% of participants
completed the postassessment (.49). The group interaction effect
was found to be nonsignificant. Additionally, no significant interaction effect was found between intervention category, participant
retention rate, and whether interventions were based on a CRF
aim/hypothesis.
Study Design
Psychosocial interventions that used a cross-sectional (CS) design were found to have a somewhat larger effect size (.47)
compared with psychosocial interventions that were based on an
AP design (.22). Similarly, a much stronger effect size was evident
for exercise interventions that were based on a CS design (.93)
compared with exercise studies that utilized an AP design (.52).
Additionally, no significant interaction effect was found between
intervention category, study design, and whether interventions
were based on a CRF aim/hypothesis.
Participant Characteristics
Cancer type. Psychosocial intervention studies that only included breast cancer patients were found to have a smaller effect
size (.22) compared with psychosocial interventions with mixed
samples (including breast cancer patients) or those explicitly focused on other non breast cancer samples (e.g., head and neck
malignancies; .55). Similarly, exercise interventions that included
mixed samples (including breast cancer patients) or explicitly
included other non breast cancer samples (.92) were found to have
a much stronger effect size than exercise studies that included only
breast cancer patients (.60). Moreover, a significant interaction
effect was found between intervention category, heterogeneous
cancer sample, and whether the intervention was based on a CRF
aim/hypothesis ( p .05). In particular, the one exercise study that
did not specify a CRF aim/hypothesis was found to have a substantially larger effect size (1.42) relative to the exercise interventions that included a CRF aim (.42). Psychosocial interventions
that included heterogeneous cancer samples were found to have
moderate effect sizes regardless of whether they included a CRF
aim/hypothesis.
Medical/cancer treatment status. Psychosocial intervention
studies that were conducted while patients were still undergoing
primary or adjuvant cancer treatment (e.g., radiation or chemotherapy) were found to have a small to moderate effect size (.27),
whereas studies that included interventions with patients after
treatment (some participants who were on treatment combined
with others who had completed their treatment) had a somewhat
larger effect size (.49). A similar pattern emerged for exercise
interventions. That is, studies that included participants not in
727
Intervention Variables
Treatment modality. The effect size for psychosocial interventions that were administered on an individual basis (.29) was
relatively small compared with effect sizes for interventions administered in a group format (.48). All of the exercise interventions
used individual-based programs, which were found to have a
moderately strong effect (.69).
Treatment setting. The effect size for psychosocial interventions administered in a hospital setting was moderately strong (.44)
compared with those administered in the home or across settings
(i.e., some sessions were conducted at the hospital while others
were conducted at home; .18). In contrast, exercise interventions
that were conducted at home or were administered in combined
settings were found to have a slightly better effect size (.74)
compared with exercise programs that were explicitly administered in a hospital setting (.65). Additionally, a significant interaction effect was found between intervention category, homebased treatments, and whether the intervention was based on a
CRF aim/hypothesis ( p .05). Exercise interventions that included a CRF aim/hypothesis and were conducted at home were
found to have a significantly larger effect size (.97) relative to
psychosocial interventions that were home-based and included a
CRF aim/hypothesis (.29). Both exercise and psychosocial studies
that were conducted primarily at home and did not include a CRF
aim/hypothesis were found to have a moderately small effect size
(.27 to .29).
Treatment duration. Psychosocial interventions that consisted
of fewer than 6 sessions were found to have a moderate effect size
(.47) compared with those containing 6 to 8 sessions (.36) as well
as those with more than 8 sessions (.30). Exercise programs with
between 6 and 8 sessions were found to have a moderately strong
effect size (.75), which was comparable to the moderately strong
effect size for programs with more than 8 sessions (.67).
728
Exercise Treatment Components
DISCUSSION
A total of 119 articles in scientific peer-reviewed journals published in the English language were identified that explored the
effectiveness of non-pharmacological interventions for fatiguerelated variables as either primary or secondary outcomes. Because
there is currently no universal gold standard measure of CRF and
the literature includes studies defining CRF as (a) a loss of energy,
feelings of tiredness, and exhaustion (including loss of focus and
concentration) and/or (b) an increase in energy, as indexed by
measures of vigor and vitality, we chose to separately assess the
efficacy of non-pharmacological interventions according to their
effect on (a) reducing fatigue/tiredness and (b) enhancing vigor/
vitality and energy levels.
CANCER-RELATED FATIGUE
729
Table 7
Psychological Interventions: Sample Size, Effect Size, Confidence Interval, and Significance for Vigor/Vitality Outcome for Each
Study Variable at Longer Term Follow-ups
Effect size (ES)
(weighted)
Variable
Psychological studies with additional follow-upsa
Psychological: Within 4 months
Psychological: 6 months
Psychological: 6 months
CRF aim
Within 4 months
6 months
6 months
Non-CRF aim
Within 4 months
6 months
6 months
Vigor/vitality specific theory
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
No vigor/vitality specific theory
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
POMS Vigor subscale
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Non-POMS vigor measure
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment modality: individual therapy
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment modality: group therapy
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Baseline compatibility and/or adjustment made
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
No. of study
trials: K
ES
p (random)
Lower
Upper
13
2
5
6
0.289
0.280
0.370
0.224
0.000
0.000
0.001
0.000
0.206
0.123
0.370
0.124
0.371
0.437
0.147
0.324
0
0
2
0.130
0.030
0.012
0.248
2
5
4
0.280
0.280
0.271
0.000
0.001
0.002
0.123
0.370
0.101
0.437
0.147
0.440
0.130
0.030
0.012
0.248
0.280
0.000
0.123
0.437
0.370
0.001
0.147
0.593
0.271
0.002
0.101
0.440
0.280
0.000
0.123
0.437
0.370
0.001
0.147
0.593
0.208
0.008
0.053
0.362
0.130
0.460
0.030
0.000
0.012
0.445
0.248
0.475
0.360
0.000
0.343
0.377
0.266
0.130
0.231
0.005
0.030
0.075
0.080
0.012
-0.024
0.452
0.248
0.486
0.200
0.000
0.179
0.221
0.440
0.049
0.001
0.879
0.310
0.039
0.016
0.604
0.280
0.000
0.123
0.437
0.370
0.130
0.271
0.001
0.030
0.002
0.147
0.012
0.101
0.593
0.248
0.440
(table continues)
0
0
0
0
2
0
0
2
0
5
0
4
0
2
0
5
0
3
0
0
0
0
2
1
0
1
0
2
2
2
0
1
0
3
0
2
0
2
0
5
2
4
730
Table 7 (continued )
No. of study
trials: K
ES
p (random)
Lower
Upper
0.360
0.000
0.343
0.377
0.605
0.130
0.410
0.014
0.030
0.000
0.125
0.012
0.312
1.085
0.248
0.508
0.200
0.000
0.179
0.221
0.212
0.000
0.136
0.289
0.135
0.000
0.077
0.193
0.360
0.000
0.343
0.377
0.503
0.003
0.174
0.833
0.327
0.001
0.129
0.525
0.200
0.000
0.179
0.221
0.170
0.130
0.100
0.000
0.030
0.000
0.152
0.012
0.051
0.188
0.248
0.149
0.360
0.000
0.343
0.377
0.360
0.000
0.343
0.377
0.231
0.075
-0.024
0.486
0.200
0.000
0.179
0.221
0.373
0.130
0.310
0.030
0.030
0.039
0.035
0.012
0.016
0.710
0.248
0.604
0.200
0.000
0.179
0.221
0.212
0.130
0.281
0.000
0.030
0.119
0.136
0.012
-0.072
0.289
0.248
0.634
0.360
0.000
0.343
0.377
0.605
0.014
0.125
1.085
0.260
0.009
0.064
0.456
0.280
0.000
0.123
0.437
0.234
0.001
0.092
0.375
0.310
0.000
0.180
0.439
0
0
1
0
2
2
2
0
1
0
3
0
2
0
1
0
3
0
3
0
1
0
2
2
1
0
1
0
1
0
2
0
1
0
4
2
2
0
1
0
3
2
2
0
1
0
2
0
2
0
2
0
3
0
3
CANCER-RELATED FATIGUE
731
Table 7 (continued )
Effect size (ES)
(weighted)
Variable
Not on treatment at initial assessment
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Good validityb
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Poor validityc
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Institutional setting
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Home or combined home and institution setting
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment duration: less than 6 sessions
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment duration: 6 to 8 sessions
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Treatment duration: 8 sessions
Within 4 months, with CRF aim
Within 4 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
6 months, with CRF aim
6 months, with non-CRF aim
Specific treatment approaches: within 4 months
CBT
Specific treatment approaches: 6 months
CBT
No. of study
trials: K
ES
p (random)
Lower
Upper
0
0
0
2
2
1
0.575
0.130
0.160
0.036
0.030
0.000
0.036
0.012
0.152
1.114
0.248
0.168
0.360
0.000
0.343
0.377
0.503
0.130
0.271
0.003
0.030
0.002
0.174
0.012
0.101
0.833
0.248
0.440
0.200
0.000
0.179
0.221
0.170
0.000
0.152
0.188
0.280
0.000
0.123
0.437
0.420
0.002
0.160
0.680
0.271
0.002
0.101
0.440
0.170
0.130
0.000
0.030
0.133
0.012
0.207
0.248
0.360
0.000
0.343
0.377
0.360
0.130
0.231
0.000
0.030
0.075
0.343
0.012
-0.024
0.377
0.248
0.486
0.200
0.000
0.179
0.221
0.440
0.049
0.001
0.879
0.160
0.000
0.152
0.168
0.170
0.000
0.133
0.207
0.460
0.000
0.445
0.475
0.280
0.000
0.123
0.437
0.370
0.001
0.147
0.593
(table continues)
0
1
0
3
2
4
0
1
0
2
0
0
0
2
0
4
0
4
0
0
0
1
2
0
0
1
0
1
2
2
0
1
0
3
0
1
0
0
0
1
0
1
732
Table 7 (continued )
No. of study
trials: K
2
0
0
2
2
ES
p (random)
Lower
Upper
0.260
0.009
0.064
0.456
0.075
0.325
0.000
0.016
0.053
0.060
0.096
0.590
Note. CRF cancer-related fatigue; ITT intention to treat; CBT cognitive behavioral therapy.
The effects from the psychological studies that included at least one additional following assessment were categorized according to short (i.e., within 4
months follow-up), medium (6-month follow-up), and longer term (more than 6 months) effects. b Good validity is represented by a score above 4 points
out of 8. c Poor validity is represented by a score of less than 4.5 points.
p .05. p .01.
a
CANCER-RELATED FATIGUE
733
734
CANCER-RELATED FATIGUE
735
Limitations
The present review has several potential limitations. First, by
excluding articles that were not published in scientific peerreviewed journals and those not printed in the English language, it
is probable that a number of relevant treatment intervention studies
were omitted. Similarly, akin to any review, although thorough
736
Summary
The present systematic evaluation of the quality and outcomes
of non-pharmacological interventions indicates that both psychosocial and exercise-based therapies show strong potential for effectively ameliorating CRF, particularly when specific therapeutic
orientations (notably, multimodal exercise and walking interventions as well as restorative, supportive expressive, and CBT psychosocial therapies) are considered. This is a noteworthy conclusion given the increasing impression in the literature in recent
years emphasizing the potential benefits of exercise-based interventions in managing CRF relative to other non-pharmacological
types of interventions (e.g., Lawrence et al., 2004; Mock, 2004;
Stasi et al., 2003). To our knowledge this is the most comprehensive review conducted to date that has systematically evaluated the
References
References marked with an asterisk indicate studies included in the
meta-analysis, and references marked with a dagger indicate studies included in the systematic review.
Adamsen, L., Midtgaard, J., Rorth, M., Borregaard, N., Andersen, C.,
Quist, M., et al. (2003). Feasibility, physical capacity, and health benefits of a multidimensional exercise program for cancer patients undergoing chemotherapy. Supportive Care in Cancer, 11, 707716.
Ahlberg, K., Ekman, T., Gaston-Johansson, F., & Mock, V. (2003, May 7).
Assessment and management of cancer-related fatigue in adults. Lancet,
362, 640 650. Retrieved from http://image.thelancet.com/extras/
02art6023web.pdf
*Ahles, T. A., Tope, D. M., Pinkson, B., Walch, S., Hann, D., Whedon,
M., et al. (1999). Massage therapy for patients undergoing autologous
bone marrow transplantation. Journal of Pain and Symptom Management, 18, 157163.
Allison, P. J., Edgar, L., Nicolau, B., Archer, J., Black, M., & Hier, M.
(2004). Results of a feasibility study for a psycho-educational intervention in head and neck cancer. Psycho-Oncology, 13, 482 485.
Allison, P. J., Nicolau, B., Edgar, L., Archer, J., Black, M., & Hier, M.
(2004). Teaching head and neck cancer patients coping strategies: Results of a feasibility study. Oral Oncology, 40, 538 544.
Altman, D. G., Schulz, K. F., Moher, D., Egger, M., Davidoff, F., Elbourne, D., et al. (2001). The revised CONSORT statement for reporting
randomized trials: Explanation and elaboration. Annals of Internal Medicine, 134, 663 694.
Andrykowski, M. A., Schmidt, J. E., Salsman, J. M., Beacham, A. O., &
Jacobsen, P. B. (2005). Use of a case definition approach to identify
cancer-related fatigue in women undergoing adjuvant therapy for breast
cancer. Journal of Clinical Oncology, 20, 23, 6613 6622.
*Arathuzik, D. (1994). Effects of cognitive behavioral strategies on pain
in cancer patients. Cancer Nursing, 17, 207214.
*Badger, T. A., Braden, C. J., & Mishel, M. H. (2001). Depression
burden, self-help interventions, and side effect experience in women
receiving treatment for breast cancer. Oncology Nursing Forum, 28,
567574.
*Badger, T., Segrin, C., Meek, P., Lopez, A., Bonham, E., & Sieger, A.
CANCER-RELATED FATIGUE
(2005). Telephone interpersonal counseling with women with breast
cancer: Symptom management and quality of life. Oncology Nursing
Forum, 32, 273279.
Bailey, L. M. (1983). The effects of live music versus tape-recorded
music on hospitalized cancer patients. Music Therapy, 3, 1728.
*Barsevick, A. M., Dudley, W., Beck, S., Sweeney, C., Whitmer, K., &
Nail, L. (2004). A randomized clinical trial of energy conservation for
patients with cancer-related fatigue. Cancer, 100, 13021310.
Barsevick, A. M., Whitmer, K., Sweeney, C., & Nail, L. M. (2002). A
pilot study examining energy conservation for cancer treatment-related
fatigue. Cancer Nursing, 25, 333341.
Berger, A. M., VonEssen, S., Kuhn, B. R. Piper, S., Agrawal, S., Lynch,
J. C., & Higginbotham, P. (2003). Adherence, sleep, and fatigue outcomes after adjuvant breast cancer chemotherapy: Results of a feasibility
intervention study. Oncology Nursing Forum, 30, 513522.
Berger, A. M., VonEssen, S., Kuhn, B. R. Piper, S. Farr, L., Agrawal, S.,
et al. (2002). Feasibility of a sleep intervention during adjuvant breast
cancer chemotherapy. Oncology Nursing Forum, 29, 14311441.
Berglund, G., Bolund, C., Gustavsson, U.-L., & Sjoden, P.-O. (1993).
Starting againA comparison study of a group rehabilitation program
for cancer patients. Acta Oncologica, 32, 1521.
Berglund, G., Bolund, C., Gustafsson, U.-L., & Sjoden, P.-O. (1994a). A
randomized study of a rehabilitation program for cancer patients: The
Starting Again group. Psycho-Oncology, 3, 109 120.
*Berglund, G., Bolund, C., Gustafsson, U.-L., & Sjoden, P.-O. (1994b).
One-year follow-up of the Starting Again group rehabilitation programme for cancer patients. European Journal of Cancer, 30A, 1744
1751.
*Boesen, E. H., Ross, L., Frederiksen, K., Thomsen, B. L., Dahlstrom, K.,
Schmidt, G., et al. (2005). Psychoeducational intervention for patients
with cutaneous malignant melanoma: A replication study. Journal of
Clinical Oncology, 23, 1270 1277.
Bordeleau, L., Szalai, J. P., Ennis, M., Leszcz, M., Speca, M., Sela, R., et
al. (2003). Quality of life in a randomized trial of group psychosocial
support in metastatic breast cancer: Overall effects of the intervention
and an exploration of missing data. Journal of Clinical Oncology, 21,
1944 1951.
Bower, J. E., Ganz, P. A., Desmond, K. A., Bernaards, C., Rowland, J. H.,
Meyerowitz, B. E., et al. (2006). Fatigue in long-term breast carcinoma
survivors a longitudinal investigation. Cancer, 106, 751758.
*Bridge, L. R., Benson, P., Pietroni, P. C., & Priest, R. G. (1988).
Relaxation and imagery in the treatment of breast cancer. British Medical Journal, 297, 1169 1172.
*Brown, P., Clark, M. M., Atherton, P., Huschka, M., Sloan, J. A.,
Gamble, G., et al. (2006). Will improvement in quality of life (QOL)
impact fatigue in patients receiving radiation therapy for advanced
cancer? American Journal of Clinical Oncology, 29, 5258.
*Burnham, T. R., & Wilcox, A. (2002). Effects of exercise on physiological and psychological variables in cancer survivors. Medicine and
Science in Sports and Exercise, 34, 18631867.
*Campbell, A., Mutrie, N., White, F., McGuire, F., & Kearney, N. (2005).
A pilot study of a supervised group exercise programme as a rehabilitation treatment for women with breast cancer receiving adjuvant treatment. European Journal of Oncology Nursing, 9, 56 63.
Carlson, L. E., & Garland, S. N. (2005). Impact of mindfulness-based
stress reduction (MBSR) on sleep, mood, stress and fatigue symptoms in
cancer outpatients. International Journal of Behavioral Medicine, 12,
278 285.
Carlson, L. E., Speca, M., Patel, K. D., & Goodey, E. (2003).
Mindfulness-based stress reduction in relation to quality of life, mood,
symptoms of stress, and immune parameters in breast and prostate
cancer outpatients. Psychosomatic Medicine, 65, 571581.
Carlson, L. E., Ursuliak, Z., Goodey, E., Angen, M., & Speca, M. (2001).
The effects of a mindfulness meditation-based stress reduction program
737
on mood and symptoms of stress in cancer outpatients: 6-month followup. Supportive Care in Cancer, 9, 112123.
Cella, D., Davis, K., Breitbart, W., & Curt, G. (2001). Cancer-related
fatigue: Prevalence of proposed diagnostic criteria in a United States
sample of cancer survivors. Journal of Clinical Oncology, 19, 3385
3391.
Cella, D., Peterman, A., Breitbart, S., & Curt, G. (1998). Progress toward
guidelines for the management of fatigue. Oncology, 12, 369 377.
Christopher, K. A., & Morrow, L. L. (2004). Evaluating a communitybased exercise program for women cancer survivors. Applied Nursing
Research, 17, 100 108.
Cimprich, B. (1993). Development of an intervention to restore attention in
cancer patients. Cancer Nursing, 16, 8392.
Cimprich, B., & Ronis, D. L. (2003). An environmental intervention to
restore attention in women with newly diagnosed breast cancer. Cancer
Nursing, 26, 284 292.
*Classen, C., Butler, L. D., Koopman, C., Miller, E., DiMiceli, S.,
Giese-Davis, J., et al. (2001). Supportive expressive group therapy and
distress in patients with metastatic breast cancer: A randomized clinical
intervention trial. Archives of General Psychiatry, 58, 494 501.
Cohen, J. (1988). Statistical power analysis for behavioral sciences (2nd
ed.). New York: Academic Press.
*Cohen, L., Warneke, C., Fouladi, R. T. Rodriguez, M. A., & ChaoulReich, A. (2004). Psychological adjustment and sleep quality in a
randomized trial of the effects of a Tibetan yoga intervention in patients
with lymphoma. Cancer, 100, 22532260.
*Coleman, E. A., Coon, S., Hall-Barrow, J., Richards, K., Gaylor, D., &
Stewart, B. (2003). Feasibility of exercise during treatment for multiple
myeloma. Cancer Nursing, 26, 410 419.
*Coleman, E. A., Hall-Barrow, J., Coon, S., & Stewart, C. B. (2003).
Facilitating exercise adherence for patients with multiple myeloma.
Clinical Journal of Oncology Nursing, 7, 529 540.
*Courneya, K. S., Friedenreich, C. M., Sela, R. A., Quinney, A., Fields,
A. L. A., Jones, L. W., et al. (2003). A randomized trial of exercise and
quality of life in colorectal cancer survivors. European Journal of
Cancer Care, 12, 347357.
Courneya, K. S., Friedenreich, C. M., Sela, R. A., Quinney, A., Rhodes,
R. E., & Handman, M. (2003). The group psychotherapy and homebased physical exercise (group-hope) trial in cancer survivors: Physical
fitness and quality of life outcomes. Psycho-Oncology, 12, 357374.
*Courneya, K. S., Mackey, J. R., Bell, G. J., Jones, L. W., Field, C. J., &
Fairey, A. S. (2003). Randomized controlled trial of exercise training in
postmenopausal breast cancer survivors: Cardiopulmonary and quality
of life outcomes. Journal of Clinical Oncology, 21, 1660 1668.
Cunningham, A. J., Edmonds, C. V. I., Jenkins, G., & Lockwood, G. A.
(1995). A randomized comparison of two forms of a brief, group,
psychoeducational program for cancer patients: Weekly sessions versus
a weekend intensive. International Journal of Psychiatry in Medicine,
25, 173189.
Cunningham, A. J., Lockwood, G. A., & Edmonds, C. V. (1993). Which
cancer patients benefit most from a brief, group, coping skills program?
International Journal of Psychiatry in Medicine, 23, 383398.
Davidson, J. R., Waisberg, J. L., Brundage, M. D., & Maclean, A. W.
(2001). Nonpharmacologic group treatment of insomnia: A preliminary
study with cancer survivors. Psycho-Oncology, 10, 389 397.
*De-Moor, C., Sterner, J., Hall, M., Warneke, C., Gilani, Z., Amato, R.,
et al. (2002). A pilot study of the effects of expressive writing on
psychological and behavioral adjustment in patients enrolled in a Phase
II trial of vaccine therapy for metastatic renal cell carcinoma. Health
Psychology, 21, 615 619.
de Wit, R., van Dam, F., Zandbelt, L., van Buuren, A., van der Heijden,
K., Leenhouts, G., et al. (1997). A pain education program for chronic
cancer pain patients: Follow-up results from a randomized controlled
trial. Pain, 73, 55 69.
738
Gruber, B. L., Hersh, S. P., Hall, N. R. S., Waletzky, L. R., Kunz, J. F.,
Carpenter, J. K., et al. (1993). Immunological responses of breast cancer
patients to behavioral interventions. Biofeedback and Self-Regulation,
18, 122.
*Headley, J. A., Ownby, K. K., & John, L. D. (2004). The effect of seated
exercise on fatigue and quality of life in women with advanced breast
cancer. Oncology Nursing Forum, 31, 977983.
*Helgeson, V. S., Cohen, S., Schulz, R., & Yasko, J. (1999). Education
and peer discussion group interventions and adjustment to breast cancer.
Archives of General Psychiatry, 56, 340 347.
*Helgeson, V. S., Cohen, S., Schulz, R., & Yasko, J. (2001). Long-term
effects of educational and peer discussion group interventions on adjustment to breast cancer. Health Psychology, 20, 387392.
Hernandez-Reif, M., Field, T., Ironson, G., Beutler, J., & Vera, Y. (2005).
Natural killer cells and lymphocytes increase in women with breast
cancer following massage therapy. International Journal of Neuroscience, 115, 495510.
*Hernandez-Reif, M., Ironson, G., Field, T., Hurley, J., Katz, G., Diego,
M., et al. (2004). Breast cancer patients have improved immune and
neuroendocrine functions following massage therapy. Journal of Psychosomatic Research, 57, 4552.
Hofmana, M., Ryan, J. L., Figueroa-Moseley, C. D., Jean-Pierre, P., &
Morrow, G. R. (2007). Cancer-related fatigue: The scale of the problem.
The Oncologist, 12, 4 10.
Holley, S., & Borger, D. (2001). Energy for living with cancer: Preliminary findings of a cancer rehabilitation group intervention study. Oncology Nursing Forum, 28, 13931396.
Hosaka, T. (1996). A pilot study of a structured psychiatric intervention
for Japanese women with breast cancer. Psycho-Oncology, 5, 59 64.
Hosaka, T., Sugiyama, Y., Hirai, K. Okuyama, Y., Sugawara, Y., &
Nakamura. (2001). Effects of a modified group intervention with
early-stage breast cancer patients. General Hospital Psychiatry, 23,
145151.
Hosaka, T., Sugiyama, Y., & Tokuda, Y. (2000). Effects of a structured
psychiatric intervention on cancer patients emotions and coping styles.
International Journal of Clinical Oncology, 5, 188 191.
Hosaka, T., Sugiyama, Y., Tokuda, Y., & Okuyama, T. (2000). Persistent
effects of a structured psychiatric intervention on breast cancer patients
emotions. Psychiatry and Clinical Neurosciences, 54, 559 563.
Hosaka, T., Sugiyama, Y., Tokuda, Y., Okuyama, T., Sugawara, Y., &
Nakamura, Y.. (2000). Persistence of the benefits of a structured psychiatric intervention for breast cancer patients with lymph node metastases. Tokai Journal of Experimental and Clinical Medicine, 25, 45 49.
Hosaka, T., Tokuda, Y., Sugiyama, Y., Hirai, K., & Okuyama, T. (2000).
Effects of a structured psychiatric intervention on immune function of
cancer patients. Journal of Experimental and Clinical Medicine, 25,
183188.
Jacobsen, P. B., Donovan, K. A., Vadaparampil, S. T., & Small, B. J.
(2007). Systematic review and meta-analysis of psychological and
activity-based interventions for cancer-related fatigue. Health Psychology, 26, 660 667.
*Jacobsen, P. B., Meade, C. D., Stein, K. D., Chirikos, T. N., Small, B. J.,
& Ruckdeschel, J. C. (2002). Efficacy and costs of two forms of stress
management training for cancer patients undergoing chemotherapy.
Journal of Clinical Oncology, 20, 28512862.
Jean-Pierre, P., Figueroa-Moseley, C. D., Kohli, S., Fiscella, K., Palesh, O. G.,
& Morrow, G. R. (2007). Assessment of cancer-related fatigue: Implications
for clinical diagnosis and treatment. The Oncologist, 12, 1121.
Kamangar, F., Dores, G. M., & Anderson, W. F. (2006). Patterns of cancer
incidence, mortality, and prevalence across five continents: Defining
priorities to reduce cancer disparities in different geographic regions of
the world. Journal of Clinical Oncology, 24, 21372150.
Knols, R., Aaronson, N. K., Uebelhart, D., Fransen, J., & Aufdemkampe,
G. (2005). Physical exercise in cancer patients during and after medical
CANCER-RELATED FATIGUE
treatment: A systematic review of randomized and controlled clinical
trials. Journal of Clinical Oncology, 23, 3830 3842.
Lawrence, D. P., Kupelnick, B., Miller, K., Devine, D., & Lau, J. (2004).
Evidence report on the occurrence, assessment, and treatment of fatigue
in cancer patients. Journal of the National Cancer Institute Monographs,
32, 40 50.
Lepore, S. J., & Helgeson, V. S. (1999). Psychoeducational support group
enhances quality of life after prostate cancer. Cancer Research Therapy
and Control, 8, 8191.
Markes, M., Brockow, T., & Resch, K. L. (2006). Exercise for women
receiving adjuvant therapy for breast cancer Cochrane Database of
Systematic Reviews, 4. Art. No. CD005001. DOI: 10.1002/
14651858.CD005001.pub2
*Maughan, K., & Clarke, C. (2001). The effect of a clinical nurse
specialist in gynaecological oncology on quality of life and sexuality.
Journal of Clinical Nursing, 10, 221229.
McKenzie, D. C., & Kalda, A. L. (2003). Effect of upper extremity
exercise on secondary lymphedema in breast cancer patients: A pilot
study. Journal of Clinical Oncology, 21, 463 466.
McNair, D. M., Lorr, M., & Droppelman, L. (1971). Profile of Mood
States: Manual. San Diego, CA: EDITS/Educational and Industrial
Testing Service Inc.
McNeely, M. L., Campbell, K. L., Rowe, B. H., Klassen, T. P., Mackey,
J. R., & Courneya, K. S. (2006). Effects of exercise on breast cancer
patients and survivors: A systematic review and meta-analysis. Canadian Medical Association Journal, 175, 34 41.
Meek, P. M., Nail, L. M., Barsevick, A., Schwartz, A. L., Stephen, S.,
Whitmer, K., et al. (2000). Psychometric testing of fatigue instruments
for use with cancer patients. Nursing Research, 49, 181190.
Mendoza, T. R., Wang, X. S., Cleeland, C. S., Morrissey, H., Johnson,
B. A., Wendt, J. K., et al. (1999). The rapid assessment of fatigue
severity in cancer patientsUse of the Brief Fatigue Inventory. Cancer,
85, 1186 1196.
Mock, V. (2001). Fatigue management evidence and guidelines for practice. Cancer, 92, 1699 1707.
Mock, V. (2004). Evidence-based treatment for cancer-related fatigue.
Journal of the National Cancer Institute Monographs, 32, 112118.
*Mock, V., Burke, M. B., Sheehan, P. Creaton, E. M., Winningham,
M. L., McKenney-Tedder, S., et al. (1994). A nursing rehabilitation
program for women with breast cancer receiving adjuvant chemotherapy. Oncology Nursing Forum, 21, 899 907.
*Mock, V., Dow, K. H., & Meares, C. J. (1997). Effects of exercise on
fatigue, physical functioning, and emotional distress during radiation
therapy for breast cancer. Oncology Nursing Forum, 24, 9911000.
*Mock, V., Frangakis, C., Davidson, N. E., Ropka, M. E., Pickett, M.,
Poniatowski, B., et al. (2005). Exercise manages fatigue during breast
cancer treatment: A randomized controlled trial. Psycho-Oncology, 14,
464 477.
*Mock, V., Pickett, M., Ropka, M. E., Lin, E. M., Stewart, K. J., Rhodes,
V. A., et al. (2001). Fatigue and quality of life outcomes of exercise
during cancer treatment. Cancer Practice, 9, 119 127.
Moher, D., Schulz, K., & Altman, D. G. (2001). The CONSORT statement:
Revised recommendations for improving the quality of reports of
parallel-group randomised trials. The Lancet, 357, 11911194.
Molassiotis, A., Yung, H. P., Yam, B. M., Chan, F. Y. S., & Mok, T. S. K.
(2002). The effectiveness of progressive muscle relaxation training in
managing chemotherapy-induced nausea and vomiting in Chinese breast
cancer patients: A randomized controlled trial. Support Care in Cancer,
10, 237246.
Mustian, K. M., Morrow, G. R., Carroll, J. K., Figueroa-Moseley, C. D.,
Jean-Pierre, P., & Williams, G. C. (2007). Integrative nonpharmacologic
behavioral interventions for the management of cancer-related fatigue.
The Oncologist, 12(Suppl. 1), 52 67.
Oldervoll, L. M., Kaasa, S., Knobel, H., & Loge, J. H. (2003). Exercise
739
740
*Sandgren, A. K., McCaul, K. D., King, B., ODonnell, S., & Foreman, G.
(2000). Telephone therapy for patients with breast cancer. Oncology Nursing Forum, 27, 683 688.
Schmitz, K. H., Holtzman, J., Courneya, K. S., Masse, L. C., Duval, S., &
Kane, R. (2005). Controlled physical activity trials in cancer survivors:
A systematic review and meta-analysis. Cancer Epidemiology Biomarkers & Prevention, 14, 1588 1595.
Schneider, S. M., Ellis, M., Coombs, W. T., Shonkwiler, E. L., & Folsom,
L. C. (2003). Virtual reality intervention for older women with breast
cancer. Cyberpsychology and Behavior, 6, 301307.
Schneider, S. M., Prince-Paul, M., Allen, M. J., Silverman, P., & Talaba,
D. (2004). Virtual reality as a distraction intervention for women receiving chemotherapy. Oncology Nursing Forum, 31, 81 88.
Schubert, C., Hong, S., Natarajan, L., Mills, P. J., & Dimsdale, J. E. (2007).
The association between fatigue and inflammatory marker levels in
cancer patients: A quantitative review. Brain, Behavior, and Immunity,
21, 413 427.
Schwartz, A. L. (1998). The Schwartz Cancer Fatigue Scale: Testing
reliability and validity. Oncology Nursing Forum, 25, 711717.
Schwartz, A. L. (1999). Fatigue mediates the effects of exercise on quality
of life. Quality of Life Research, 8, 529 538.
Schwartz, A. L. (2000a). Daily fatigue patterns and effect of exercise in
women with breast cancer. Cancer Practice, 8, 16 24.
Schwartz, A. L. (2000b). Exercise and weight gain in breast cancer
patients receiving chemotherapy. Cancer Practice, 8, 231237.
Schwartz, A. L., Mori, M., Gao, R., Nail, L. M., & King, M. E. (2001).
Exercise reduces daily fatigue in women with breast cancer receiving
chemotherapy. Medicine Science Sports and Exercise, 33, 718 723.
Schwartz, J. E., Jandorf, L., & Krugg, L. B. (1993). The measurement of
fatigue: A new instrument. Journal of Psychosomatic Research, 37,
753762.
*Segal, R., Evans, W., Johnson, D. Smith, J., Colletta, S., Gayton, J., et
al. (2001). Structured exercise improves physical functioning in women
with Stages I and II breast cancer: Results of a randomized controlled
trial. Journal of Clinical Oncology, 19, 657 665.
*Segal, R. J., Reid, R. D., & Courneya, K. S. (2003). Resistance exercise
in men receiving androgen deprivation therapy for prostate cancer.
Journal of Clinical Oncology, 21, 16531659.
Smith, M. L., Glass, G. V., & Miller, T. I. (1980). The benefits of
psychotherapy. Baltimore, MD: John Hopkins University Press.
*Speca, M., Carlson, L. E., Goodey, E., & Angen, M. (2000). A randomized, wait-list controlled clinical trial: The effect of a mindfulness
meditation-based stress reduction program on mood and symptom of
stress in cancer outpatients. Psychosomatic Medicine, 62, 613 622.
Spiegel, D., & Bloom, J. R. (1983). Group therapy and hypnosis reduce
metastatic breast carcinoma pain. Psychosomatic Medicine, 45, 333339.
*Spiegel, D., Bloom, J. R., & Yalom, I. (1981). Group support for patients
with metastatic cancer: A randomized outcome study. Archives of General Psychiatry, 38, 527533.
*Stanton, A. L., Ganz, P. A., Kwan, L., Meyerowitz, B. E., Bower, J. E.,
Krupnick, J. L., et al. (2005). Outcomes from the Moving Beyond
Cancer psychoeducational, randomized, controlled trial with breast cancer patients. Journal of Clinical Oncology, 23, 6009 6018.
Stasi, R., Abriani, L., Beccaglia, P., Terzoli, E., & Amadori, S. (2003).
Cancer-related fatigue evolving concepts in evaluation and treatment.
Cancer, 98, 1786 1801.
Stevinson, C., Lawlor, D. A., & Fox, K. R. (2004). Exercise interventions
for cancer patients: Systematic review of controlled trials. Cancer
Causes and Control, 15, 10351056.
Stone, P. (2002). The measurement, causes and effective management of
cancer-related fatigue. International Journal of Palliative Nursing, 8,
120 128.
Stone, P., Richards, M., & Hardy, J. (1998). Fatigue in patients with
cancer. European Journal of Cancer, 34, 1670 1676.
CANCER-RELATED FATIGUE
Wu, H., & McSweeney, M. (2001). Measurement of fatigue in people with
cancer. Oncology Nursing Forum, 28, 13711384.
*Wydra, W. (2001). The effectiveness of a self-care management
interactive multimedia module. Oncology Nursing Forum, 28, 1399
1407.
Yellen, S. B., Cella, D. F., Webster, K., Blendowski, C., & Kaplan, E.
(1997). Measuring fatigue and other anemia-related symptoms with the
Functional Assessment of Cancer Therapy (FACT) measurement system. Journal of Pain and Symptom Management, 13, 6374.
741
M
(weighted)
95% confidence
interval
I (%)
ra
Re
Rs
GRe
ra
G
Re
Rs
C
GRe
GRs
GRera
.56
.80
.81
.80
.04
.65
.66
.10
.53.59
.76.83
.79.84
.79.82
.02.06
.61.68
.63.69
.09.11
249
236
246
237
204
236
236
236
17,319*
19,829*
10,706*
5,644*
6,249*
20,668*
17,469*
2,461*
99
99
98
96
97
99
99
90
.057
.067
.035
.019
.023
.070
.060
.008
.78**
.43**
.56**
.23**
.91**
.83**
.10
.43**
.03
.82**
.92**
.14*
.23**
.63**
.12
.05
.41**
.72**
.08
.03
.78**
Note. See The Mathematical Formulation of Brunswiks Lens Model section of the text for a description of the lens model indices. Q
represents within-group heterogeneity; I2 is the percentage of variation attributable to between-study heterogeneity; 2 is the DerSimonian and
Laird (1986) estimate of between-study variance.
*
p .05. **p .01.
DOI: 10.1037/a0013550